Conclusions

We showed that it is possible to reduce the number of species and reactions in the UMIST $95$ chemical database those are needed to compute abundances of e- and CO accurately under conditions, typical of molecular cloud cores. We found that for the pure gas-phase chemistry, the number of species can be reduced by a factor of $3$-$6$ for e- and $3$-$50$ for CO, depending on the gas density. The corresponding computational speed gain may reach factors of hundreds for e- and thousands for CO. On the contrary, for the gas-grain chemistry involving dust surface reactions, we did not succeed in isolation of a significantly reduced set of chemical species in order to speed up the calculations.